physlite-basics

name: physlite-basics description: Use when the user needs to OPEN an ATLAS DAOD_PHYSLITE file with uproot / awkward, list its branches, reconstruct calibrated objects (electrons, muons, photons, taus, small-R jets, large-R jets, MET, trigger decisions), or apply MC normalisation. Assumes a public HTTPS or XRootD URI such as root://eospublic.cern.ch/... from atlas_get_urls or cod_list_files. Does NOT cover the older flat 13 TeV reduced ntuples (use the matching tutorial via atlas-notebooks), CMS NanoAOD or MiniAOD, or sample discovery (use atlas-opendata). Disambiguator phrase: PHYSLITE branch reader. data_scope: open experiment: atlas

Scope

Use this skill when the user is handling DAOD_PHYSLITE files (the default "physics-lite" derivation format of the 2024+ ATLAS Open Data releases). For the older 13 TeV reduced ntuples used by the uproot_python tutorials in atlas-outreach-data-tools/notebooks-collection-opendata, this skill is not the right one — those are flat ntuples with bespoke branches; point users at the relevant notebook directly via atlas-notebooks.

If the user needs a DSID or cross-section, use atlas-opendata. If the user needs a portal file URI, use cern-opendata.

What PHYSLITE is

• ATLAS's official slimmed-down derivation designed for analysis without the full xAOD framework.
• Ships calibrated analysis-level objects: electrons, muons, photons, taus, small-R jets, large-R jets, MET, trigger decisions, event info.
• Readable with uproot (or xAOD via Athena; Open Data users should prefer uproot).
• Companion tutorial: for-research/physlite_tutorial.ipynb in atlas-outreach-data-tools/notebooks-collection-opendata.

Reading a file

Given a URI from atlas_get_urls(protocol='https') or from cod_list_files:

import uproot
f = uproot.open("https://opendata.cern.ch/record/.../DAOD_PHYSLITE.root")
t = f["CollectionTree"]

The interesting physics objects live under AnalysisElectronsAuxDyn, AnalysisMuonsAuxDyn, AnalysisJetsAuxDyn, AnalysisPhotonsAuxDyn, AnalysisTauJetsAuxDyn, MET_Core_AnalysisMETAuxDyn, with kinematics in pt, eta, phi, m (in MeV — watch the units).

Common branches

Units are MeV throughout PHYSLITE. Convert to GeV before plotting.

MC normalisation

Use the formula (always — do not hide it in a helper):

weight = cross_section_pb * 1000 * kFactor * genFiltEff * mcWeight / sumOfWeights * luminosity_fb

Get the per-sample terms from the atlasopenmagic MCP:

• atlas_get_metadata(dsid) → cross_section_pb, kFactor, genFiltEff, sumOfWeights.
• mcWeight is per-event, from EventInfoAuxDyn.mcEventWeights[0].
• luminosity_fb comes from the release (e.g. 36.1 fb⁻¹ for the old Hyy tutorial, 140 fb⁻¹ for Run 2 full).

Procedure

1. Get the URI for the sample with atlas_get_urls(dsid, protocol='https') (or protocol='root' for XRootD).
2. Open with uproot.open.
3. Read the branches you need with t.arrays([...], library='ak').
4. Convert MeV → GeV.
5. Compute the per-event weight using the formula above.
6. Fill histograms with hist / boost_histogram, plot with mplhep.style.ATLAS.

Pitfalls

• Units are MeV, not GeV. Most published plots divide by 1000.
• mcEventWeights is a vector, nominal is index [0]. The rest are systematic variations.
• sumOfWeights is a dataset-level quantity. Read it once from atlas_get_metadata, not per-event.
• Data files don't have mcEventWeights; guard with an is_data flag.
• For trigger decisions you need TrigMatch info which is not in PHYSLITE auxiliaries — fall back to the reduced 13 TeV ntuples if the user wants detailed trigger studies.

Verification

A successful use of this skill ends with the user having:

1. Opened a PHYSLITE file with uproot.open.
2. Extracted at least one kinematic array in GeV.
3. Printed (or plotted) a weighted distribution using the normalisation formula above.